Digital image processing method for images with bimodal tone value distribution

Claim 2

A digital image processing method for processing a digital image having shadow portions and highlight portions, said method employing a tone reproduction function generated by normalizing a sample of tone values from the digital image, said tone reproduction function being applied to tone values of said digital image to produce processed tone values, characterized by the steps of: testing said sample of tone values for bimodal distribution; if the sample is bimodal, testing said sample of tone values to determine whether the image contains useful information in the shadow portions, and if useful information is contained in the shadow portions, adjusting the contrast of the image by increasing the contrast of the shadow portions of the image and decreasing the contrast of the highlight portions of the image.

Claim 3

The digital image processing method claimed in claim 1, further characterized by the steps of: dividing the sample of tone values in two at the median tone value of the sample to form two new tone value distributions representing the highlight and shadow portions of the digital image; calculating the mean tone value of the sample of tone values and the mean tone value of the new distribution of tone values representing the highlight portions of the digital image; adding a bias to the processed tone values, said bias being proportional to the difference between the mean tone value of the new distribution of tone values representing the highlight portions of the image and the mean tone value of the sample of tone values.

Claim 4

The digital image processing method claimed in claim 1, further characterized by the steps of: dividing the sample of tone values in two at the median tone value of the sample to form two new tone value distributions representing the highlight and shadow portions of the digital image; testing the new tone value distribution representing the shadow portions of the image for skewness, and adjusting the contrast of the image only if the skewness does not exceed a predetermined amount.

Claim 5

The digital image processing method claimed in claim 1, wherein the step of testing for bimodality comprises testing for the satisfaction of the inequality 0.0385β.sub.1.sup.2 +0.9217β.sub.1 +1.7662≧β.sub.2 ≧(β.sub.1 +1.0), where β 1 and β 2 are the coefficients of skewness and peakedness, respectively, of the sample of tone values.

Claim 6

The invention claimed in claim 1, wherein the tone reproduction function relates tone values of the digital image to values of a normal variate Z, and wherein the step of adjusting the contrast of the image includes the steps of: forming a first estimate of the spread of the tone values in the shadow portions of the image and a part or all of the highlight portions of the image; forming a second estimate of the spread of the tone values in only the shadow portions of the image; forming a ratio of said first and second estimates; multiplying the Z values of the tone reproduction function below Z=0 by the ratio; and multiplying the Z values of the tone reproduction function above Z=0 by the reciprocal of the ratio.

Claim 7

A digital color image processing method for correcting the tone scale and color balance of a digital color image having three color components and having shadow portions and highlight portions, comprising the steps of: collecting a sample of color values from the informational portion of the digital color image for each of the three color components of the image; forming a color reproduction function for each of the three color components by normalizing the respective samples of color values; applying the color reproduction functions to color values of the respective color components of the digital color image to produce a processed digital color image; and testing the samples of color values for bimodal distribution, and if the sample for each color component is bimodal, testing one of the samples of color values to determine whether the digital color image contains useful image detail in the shadow portions, and if useful detail is contained in the shadow portions, adjusting the contrast of the digital color image by increasing the contrast of the shadow portions of the image and decreasing the contrast of the highlight portions of the image.

Claim 8

A digital color image processing method for correcting the tone scale and color balance of a digital color image having three color components and having shadow portions and highlight portions, comprising the steps of: collecting a sample of color values from the informational portion of the digital color image for each of the three color components of the image; forming a color reproduction function for each of the three color components by normalizing the respective samples of color values; applying the color reproduction functions to color values of the respective color components of the digital color image to produce a processed digital color image; and testing the samples of color values for bimodal distribution, and if the sample for each color component is bimodal, testing one of the samples of color values to determine whether the digital color image contains useful image detail in the shadow portions, and if useful detail is contained in the shadow portions, adjusting the contrast of the digital color image by increasing the contrast of the shadow portions of the image and decreasing the contrast of the highlight portions of the image.

Claim 9

The digital color image processing method claimed in claim 6, further characterized by the steps of: dividing one of the samples of color values in two at the median color value of the sample to form two new color value distributions representing the highlight and shadow portions of the digital color image; and testing the new color value distribution representing the shadow portions of the digital color image for skewness and adjusting the contrast of the digital color image only if the skewness is less than a predetermined amount.

Claim 10

The digital color image processing method claimed in claim 6, further comprising the steps of: dividing one of the samples of color values in two at the median color value of the sample to form two new color value distributions representing the highlight and shadow portions of the digital color image; calculating the mean tone value of said one sample of color values and the mean tone value of said new color value distribution representing the highlight portions of the digital color image; and adding a tone bias to the processed digital color image, said bias being proportional to the difference between the mean color value of the new color value distribution representing the highlight portions of the digital color image and the mean color value of said one sample of color values.

Claim 11

The invention claimed in claim 6, wherein the step of testing for bimodality comprises testing for the satisfaction of the inequality 0.0385β.sub.1.sup.2 +0.9217β.sub.1 +1.7662≧β.sub.2 ≧(β.sub.1 +1.0), where β 1 and β 2 are the coefficients of skewness and peakedness, respectively, of the samples of color values.

Claim 12

The digital color image processing method claimed in claim 6, wherein the color reproduction functions relate color values of the color components to values of a normal variate Z, and wherein the step of adjusting the contrast of the color image includes the steps of: forming a first estimate of the spread of the color values in one of the samples of color values in the shadow portions of the digital color image and a part or all of the highlight portions of the digital color image; forming a second estimate of the spread of the color values in said one sample in only the shadow portions of the digital color image; forming a ratio of said first and second estimates; multiplying the Z values of the color reproduction functions below Z=0 by the ratio; and multiplying the Z values of the color reproduction functions above Z=0 by the reciprocal of the ratio.

Claim 13

A digital color image processing method for correcting the tone scale and color balance of a digital color image having shadow portions and highlight portions comprising the steps of: (1) selecting samples of red, green and blue color values by separating red, green and blue color values of the digital color image into samples corresponding to a plurality of contrast intervals, and selecting samples of red, green and blue color values corresponding to one contrast interval based upon the similarity of the shapes of the distributions of the red, green and blue color values in the samples corresponding to the one contrast interval; (2) normalizing the selected samples of red, green and blue color values to produce red, green and blue color reproduction functions expressed in terms of values of a normal variate Z; (3) testing the selected samples of red, green and blue color values for bimodality; (4) if all three of the selected samples of color values are bimodal, testing statistical parameters of the sample of green color values to determine whether the digital color image contains useful information in the shadow portions, and if useful information is present, adjusting the contrast of the digital color image by: a. forming a first estimate of the spread of the color values in the selected sample of green color values by fitting, by linear regression, a first straight line to the entire green color reproduction function, the slope of the first straight line being the first estimate of the spread; b. forming a second estimate of the spread of the color values in the selected sample of green color values corresponding to only the shadow portions of the digital color image by fitting, by linear regression, a second straight line to the portion of the green color reproduction function corresponding to Z values between -2.0 and -0.5, the slope of the second straight line being the second estimate of the spread; c. forming a ratio of the first and second estimates; d. testing the ratio against a predetermined constant to determine whether the image contains useful information in the shadow portions, and if useful information is present: (i) multiplying the Z values of all three color reproduction functions below Z=0 by the ratio; (ii) multiplying the Z values of all three color reproduction functions above Z=0 by the reciprocal the ratio; and (iii) applying said color reproduction functions to the color values of the digital color image to produce processed color values.

Claim 14

A digital color image processing method for correcting the tone scale and color balance of a digital color image having shadow portions and highlight portions comprising the steps of: (1) selecting samples of red, green and blue color values by separating red, green and blue color values of the digital color image into samples corresponding to a plurality of contrast intervals, and selecting samples of red, green and blue color values corresponding to one contrast interval based upon the similarity of the shapes of the distributions of the red, green and blue color values in the samples corresponding to the one contrast interval; (2) normalizing the selected samples of red, green and blue color values to produce red, green and blue color reproduction functions expressed in terms of values of a normal variate Z; (3) testing the selected samples of red, green and blue color values for bimodality; (4) if all three of the selected samples of color values are bimodal, testing statistical parameters of the sample of green color values to determine whether the digital color image contains useful information in the shadow portions, and if useful information is present, adjusting the contrast of the digital color image by: a. forming a first estimate of the spread of the color values in the selected sample of green color values by fitting, by linear regression, a first straight line to the entire green color reproduction function, the slope of the first straight line being the first estimate of the spread; b. forming a second estimate of the spread of the color values in the selected sample of green color values corresponding to only the shadow portions of the digital color image by fitting, by linear regression, a second straight line to the portion of the green color reproduction function corresponding to Z values between -2.0 and -0.5, the slope of the second straight line being the second estimate of the spread; c. forming a ratio of the first and second estimates; d. testing the ratio against a predetermined constant to determine whether the image contains useful information in the shadow portions, and if useful information is present: (i) multiplying the Z values of all three color reproduction functions below Z=0 by the ratio; (ii) multiplying the Z values of all three color reproduction functions above Z=0 by the reciprocal the ratio; and (iii) applying said color reproduction functions to the color values of the digital color image to produce processed color values.

Claim 15

The digital color image processing method claimed in claim 11, further comprising the steps of dividing the selected sample of green color values in two at the median color value of the sample to form two new green color value distributions approximately representing the shadow portions and the highlight portions of the digital color image, and testing the skewness of the new green color value distribution representing the shadow portions of the digital color image and adjusting the contrast of the digital color image only if the skewness is less than a predetermined amount.

Claim 16

The digital color image processing method claimed in claim 12, further comprising the steps of: dividing the selected sample of green color values in two at the median color value of the sample to form two new green color value distributions representing the highlight and shadow portions of the digital color image; calculating the mean color values of the selected sample of green color values and the new green color value distribution representing the highlight portions of the digital color image; and adding a bias to the processed color values of the digital color image, said bias being proportional to the difference between the mean color value of the new green color value distribution representing the highlight portions of the image and the mean color value of the selected sample of green color values.